ZoneMinder had its place
ratatine last edited by
Some thoughts on this topic as I know you're trying to overcome zoneminder's limitations (and there are many.) I should note that it was built a long time ago and based on a gap in technology. Nothing like it existed for CCTV. All you had were old clunky DVRs. Browsers didn't even have good video support so they used a image-storage technique instead of video. Even now their h264 support for storage is wonky and their mobile support is meh. It needs a complete overhaul and you can't demand that of volunteers.
That said, I think you're moving in the right direction here. ZoneMinder's strengths were in its support for a variety of cameras, flexible storage, and (after sometime) an API. The API is going to be super important for the future because I think a lot of what we will do next with cameras involves computer vision.
For example, I have a Rachio controller. Imagine a situation where my cameras record a deer in my yard walking towards my petunias. It so happens that the NVR could be connected to stringify who says "Oh look! Deer. Hey Rachio... turn on the sprinklers." Petunias-1. Deer-0.
Of course this is only one silly idea and there are a lot of practical use cases. "Hey look, someone is approaching the backdoor of the house and the home alarm is turned on. You might want to take a look."
Also the old school ways of doing video surveillance are changing. We say things like "oh 10fps is good enough". Or "You want to have your camera 2' higher than the door frame." No... I want full framerate video that shows me faces - license plates - the text message they are reading on their phone! Really, if I have to go back and look I want to see that. So alas, that's a lot of zoneminder to bite off given its current design architecture.
It had its place though!
I believe ZoneMinder's strength came from the fact it was a open-source community project. Not a close-sourced platform for people to make money off every user. Personally I put up with that horror because it was free and open-source. I am not able to pay for crazily priced platforms nor do i wanna install windows for my servers. I struggled because it was my only option.
I accept ZoneMinder is old and needs a lot of work. I don't accept that the devs claim ZoneMinder is "A full-featured, open source, state-of-the-art video surveillance software system" while it remains nearly the same thing since they picked it up 4+ years ago. I understand they are volunteers and have lives to live... but I am doing the same as them. They even have more incentive than I do, they get donations and bounty support regularly. I wouldn't be surprised if their position on the team has earned them alternate sources of income. The name "ZoneMinder" gives them that luxury.
In short : ZoneMinder is dying (or dead) because it has no guardian to take care of it. It just has people editing it. ZoneMinder would still have a place if there was more effort put into the platform.. and in no way do i intend undermine what they have done so far but for the times we live in it isn't enough.
Enough of my rambling! Yes the world is evolving fast. We have to keep up or we get left behind. Luckily i have been doing a lot of testing with [email protected] for the last few months and Shinobi seems to handle it just fine. Since Shinobi uses FFmpeg you can do things like pass-through (copy encoder). Which will record or stream H.264 data with very little overhead. We can also make use of the GPU to get even better performance.
I will soon be updating object detection docs to include the fast install for OpenCV and CUDA. So you can get that detection stuff going easier.
ZM can still have a place. They just need to earn it all over again.
ratatine last edited by
Agreed on all points. The commercial solutions are horrible. I bought a nice NVR to try off Amazon and it was complete garbage. NOTHING worked and I mean nothing. They actually told me to try using an older version of the browser, complete with security holes. It was horrible and Amcrest should be ashamed of themselves.
But I digress... I've been using openRTSP to save my stream to flat files and that does a great job on one objective. I'm using Shinobi now with full frame 1920x1080 with no problems. I haven't gotten motion detection working but after I futz a bit more I'll post about that on the right thread.
The cuda integration will push me to pickup another video card I suppose (I currently run this on a VM with no GPU) but that's just the reality of doing anything like YOLO.
rockedge last edited by rockedge
@moeiscool After long term testing..and running Shinobi since it's inception (I have versions running that did nothing but stream somewhat) I have found shinobi just as prone to failure and as buggy as ZM. I really do not understand you talking about arrogance when you are exactly the same if not more boastful. Shinobi is okay but not quite the Cat's Meow. And I take direct insult as I spend a lot of time working on ZM and as you know some suggestions thrown your way. If you remember you questioned the importance of filter functions. Shinobi is a GUI for FFMEG and combines some neat features when you come right down to it. The concept ZM uses is different from Shinobi and that's it. 2 different approaches...both have some big cons and both have their strengths. What mobile support does Shinobi have actually? Because the web interface display is horrible on my phone. I don't know why you seem bent on holding grudges and resorting to bad mouthing other developers.....it makes you seem like..really young and living in mom's basement still. Some kind of coding hooligan. I seem to remember pointing that out 1.5 years ago.